Objective
The purpose of this study was to determine the rate and extent of rivaroxaban transfer across the term human placenta and determine whether passive diffusion was the primary mechanism involved in this transfer.
Study Design
The transplacental pharmacokinetics of rivaroxaban was determined with the ex-vivo placenta perfusion model. Rivaroxaban was added to the maternal or fetal circulation only (250 ng/mL). Additional experiments were conducted under equilibrative conditions with the addition of rivaroxaban to both the maternal and fetal circulations (250 ng/mL). Rivaroxaban concentrations were measured with the use of liquid chromatography–tandem mass spectrometry.
Results
There was rapid transfer of rivaroxaban across the human placenta in both the maternal-to-fetal and fetal-to-maternal directions, as evidenced by transfer ratios of 0.69 (interquartile range, 0.58–0.73; n = 5) and 0.69 (interquartile range, 0.67–0.71; n = 2), respectively, after 3 hours. Under equilibrative conditions (n = 2), rivaroxaban concentrations remained relatively constant, which suggests that rivaroxaban crosses the placenta down a concentration gradient.
Conclusion
This is the first direct evidence of rivaroxaban transfer across the term human placenta from both the mother-to-fetus and fetus-to-mother directions. Our results document that unbound rivaroxaban rapidly crosses the placental barrier via passive diffusion. However, because rivaroxaban is highly bound to plasma proteins (up to 95%), this suggests that the amount of unbound drug that may reach the fetus is likely much lower. Additional studies will need to explore its safety before administering rivaroxaban to a pregnant woman.
Anticoagulant therapy is required in some high-risk pregnancies for the prevention and treatment of venous thromboembolism (VTE) in women at high or intermediate risk for thromboembolic disease, for the prevention and treatment of systemic embolism in women with mechanical heart valves, and for the prevention of pregnancy-related complications in women with antiphospholipid antibody syndrome. A retrospective cohort study reported that VTE affected 28 per 100,000 pregnancies before birth and 65 per 100,000 after birth. Although current guidelines recommend the use of unfractionated or low molecular weight heparins, a recent open-label, randomized trial that compared antepartum prophylactic dalteparin (low molecular weight heparins) with no dalteparin found that dalteparin did not reduce the occurrence of VTE, pregnancy loss, or placenta-mediated pregnancy complications, which included severe preeclampsia, small-for-gestational-age infants, and placental abruption in pregnant women with thrombophilia who were at high risk of these complications. Similar to other drugs that are prescribed during pregnancy, the use of anticoagulant therapy can be challenging because of the possibility for both maternal and fetal complications.
Rivaroxaban (Xarelto; Bayer HealthCare AG, Levekusen, Germany) is a new oral anticoagulant, with low solubility and high membrane permeability, which increasingly is being prescribed to treat deep vein thrombosis and pulmonary embolism. After oral administration, rivaroxaban is absorbed rapidly and remains unchanged in plasma with no major or active circulating metabolites. Rivaroxaban directly inhibits Factor Xa, which is an important component of both the intrinsic and extrinsic pathways of the coagulation cascade. Because factor Xa catalyzes the conversion of prothrombin (Factor II) to thrombin (Factor IIa), it is an ideal target for anticoagulation therapy.
The objectives of our study were to determine its transplacental pharmacokinetics at term to estimate fetal drug exposure and to determine whether passive diffusion is the primary mechanism that is involved in rivaroxaban transfer across the human placenta.
Materials and Methods
Ex vivo perfusion of human placental cotyledon
The dual perfusion of a single placental cotyledon was described by Miller et al and adapted for use in our laboratory. The study was approved by the research ethics board at St Michael’s Hospital in Toronto, Ontario, Canada, and the mothers gave written consent before delivery. Placentae were obtained after elective cesarean or vaginal delivery of healthy term pregnancies (>38 weeks of gestation). The primary indications for cesarean delivery were breech position and repeat cesarean delivery. Placentae were obtained from singleton pregnancies and had to be free of macroscopic placental lesions. Women were excluded if they were taking any medication during the pregnancy, if they had any disease during pregnancy (including diabetes mellitus, hypertension, thyroid dysfunction, cancer, transplant), if they had any infection (including chorioamnionitis and TORCH infections), or if they had a positive anti–human immunodeficiency virus, anti–hepatitis B virus, or anti–hepatitis C virus serology.
Immediately after delivery, placentae were transported to the on-site laboratory in ice-cold heparinized phosphate-buffered saline solution and examined for evidence of physical damage during delivery. A well-defined artery-vein pair on the fetal side with minimal branching was chosen for cannulation, and independent maternal and fetal circulations were established within 30 minutes of delivery. Maternal and fetal blood flow rates were maintained at 13-15 and 2-3 mL/min, respectively; the temperature of the circuits and perfusion chamber was kept at 37°C. A single perfusion experiment consisted of a 1-hour preexperimental control phase followed by a 3-hour experimental phase.
Placenta perfusate consisted of 10.9 g/L M199 tissue culture medium (Sigma Aldrich, St. Louis, MO) that contained 40,000 molecular-weight dextran (maternal, 7.5 g/L; fetal, 30.0 g/L), glucose (1.0 g/L), heparin (2000 U/L), and kanamycin (100 mg/L). Antipyrine (1 mmol/L) was added to the maternal perfusate as a marker of blood flow-mediated passive diffusion. Maternal and fetal perfusates were buffered to pH 7.4 and 7.35, respectively, by the addition of small amounts of sodium bicarbonate and hydrochloric acid. The maternal perfusate was equilibrated with 95% O 2 /5% CO 2 ; the fetal perfusate was equilibrated with 95% N 2 /5% CO 2.
Preexperimental control period
The fetal and maternal circulations were maintained until all residual blood was cleared out of the vessels. The maternal and fetal circuits were then closed and replaced with 250 mL and 150 mL of fresh perfusate, respectively. O 2 and CO 2 content, pH, glucose consumption, and lactate production were measured every 15 minutes with the use of an on-site blood gas analyzer (ABL 725; Radiometer, Copenhagen, Denmark). Additional samples were taken every 15 minutes for analysis of human chorionic gonadotropin (hCG) secretion and antipyrine transfer. Tissue viability measures were calculated as previously described. Fetal arterial inflow pressure was measured in the fetal circuit proximal to its insertion into the cannulated chorionic artery, and fetal reservoir volume was measure via volume striations on the fetal reservoir vessel. An experiment was stopped if there was a loss in fetal reservoir volume >4 mL/hr or if fetal arterial inflow pressure deviated from 30-60 mm Hg for an extended period of time. At the end of the preexperimental control period, the maternal and fetal reservoirs were replaced with fresh perfusate, and the circulations were closed and recirculated.
Experimental period
Rivaroxaban was added to the maternal or fetal reservoir only or simultaneously to both the maternal and fetal reservoirs at a therapeutic concentration of 250 ng/mL at the start of the experimental period. Samples were collected from the maternal and fetal reservoirs for analysis of rivaroxaban, O 2 and CO 2 content, pH, glucose consumption, and lactate production at 0, 10, 20, 30, 60, 90, 120, 150, and 180 minutes. Additional samples were taken directly from the maternal and fetal reservoirs every 30 minutes for analysis of hCG secretion and antipyrine transfer.
Sample analysis
Perfusate samples were stored at –20°C until analysis. Antipyrine concentrations were analyzed using a ultraviolet-visible recording spectrophotometer W-160 (Shimadzu, Tokyo, Japan) at 350 nm. HCG levels were determined with the use of an enzyme-linked immunosorbent assay kit (Alpha Diagnostic International, San Antonio, TX) and a Biotek Synergy HT microplate reader (Biotek instruments, Winooski, VT) at 450 nm.
Rivaroxaban analysis was performed by the Analytical Facility for Bioactive Molecules of the Centre for the Study of Complex Childhood Diseases at the Hospital for Sick Children, Toronto, Canada. A standard curve ranging from 0–500 ng/mL for rivaroxaban (Cedarlane, Burlington, ON, Canada) was prepared in maternal or fetal perfusate. Mobile phases were (phase A) 90% water and 10% acetonitrile and (phase B) 10% water and 90% acetonitrile, with both of which consisted of 5 mmol/L ammonium formate (pH 3.2). All samples and standards were vortexed and centrifuged at 13,000 g for 5 minutes at 4°C. Fifty microliters of standards or samples were added to 450 μL of the mobile phase A that contained 10 ng/mL internal standard rivaroxaban-d4 (CacheSyn, Mississauga, Ontario, Canada). Sample analysis was performed on an AB Sciex QTrap 5500 (SCIEX, Concord, Ontario, Canada) and Agilent 1290 (Agilent Technologies Canada Inc, Mississauga, Ontario, Canada) ultra-high performance liquid chromatography system. Two microliters of sample or standard were injected at a blood flow rate of 500 μL/min through a Kinetex XB-C18 column (50 × 3.0 mm; 2.6 μm; Phenomenex, Torrance, CA) isocratically in 1.6 minutes with the use of a mobile phase composition of 60% for phase A and 40% for phase B.
The mass spectrometer was operated in positive electrospray ionization mode with a source temperature of 600°C and an internal standard voltage setting of 5500. Precursor-to-product ion mass transitions were established by standard infusions. Rivaroxaban concentrations were acquired by multiple reaction monitoring mode using transitions 436.1→145 m/z and 436.1→231.1 m/z for rivaroxaban and 440.2→145 m/z and 440.2→235.25 m/z for rivaroxaban-d4. The range of rivaroxaban quantification was 0.5-500 ng/mL. Data analysis and peak integration were performed with Analyst software (version 1.6; SCIEX). Sample concentrations were calculated by plotting peak area ratios (analyte/internal standard) against calibration curves of extracted matrix–spiked standards.
With the use of the area under the curve of rivaroxaban from the placenta perfusions, physiologic rivaroxaban concentrations were simulated with 95% protein binding to estimate maternal and fetal exposure.
Statistical analysis
All data are presented as median and interquartile range, unless stated otherwise; comparisons between preexperimental and experimental phases were analyzed with the use of a Wilcoxon sign rank test for nonnormal data. Any probability value < .05 was considered to be significant.
Results
A total of 9 cotyledons from different placentae were perfused with rivaroxaban; the physical parameters for the perfusions are shown in the Table . The mean weight of the perfused cotyledons was 23.2 ± 7.9 g. Maternal and fetal blood flow rates were 14.1 ± 0.2 and 2.4 ± 0.3 mL/min, respectively. Measures of placental viability, integrity, and function remained within normal ranges and were not significantly different between the preexperimental and experimental periods ( Table ). Lactate production remained constant for the duration of the perfusion. For all perfusions, antipyrine equalized between the maternal and fetal reservoirs after 3 hours with a final fetal-to-maternal ratio of 0.84 (interquartile range [IQR], 0.76–0.88), which is comparable with previous perfusion experiments. The rate of antipyrine disappearance from the maternal circulation was equal to the rate of appearance in the fetal circulation: 0.025 (IQR, 0.015–0.031) vs 0.023 (IQR, 0.014–0.029) μmol/g/min ( P = .77). During all perfusions, fetal reservoir volume loss was never >4 mL/hour, and pH values did not deviate from physiologic ranges.
| Viability parameter | Period, median (interquartile range) | P value | |
|---|---|---|---|
| Preexperimental | Experimental | ||
| Fetal arterial inflow pressure, mm Hg | 41.2 (38.2–48.0) | 43.7 (39.3–45.4) | .57 |
| Human chorionic gonadotropin production, mIU/g/min | 132.7 (41.4–155.1) | 116.6 (35.7–122.7) | .09 |
| Glucose consumption, μmol/g/min | 0.25 (0.18–0.25) | 0.14 (0.13–0.16) | .12 |
| Oxygen, nmol/g/min | |||
| Transfer | 4.82 (2.98–6.87) | 5.67 (3.27–7.05) | .38 |
| Delivery | 193.6 (124.89–199.46) | 191.38 (123.80–195.68) | .89 |
| Consumption | 16.08 (14.12–28.82) | 14.97 (12.85–22.78) | .30 |
After the addition of rivaroxaban (250 ng/mL) to the maternal reservoir, there was rapid transfer of rivaroxaban from the maternal-to-fetal circulations ( Figure 1 ). There was a biphasic decline in rivaroxaban concentrations in the maternal circulation, as characterized by a rapid decline in the first 30 minutes and a slower decline for the remaining 150 minutes (–2.35 vs –0.19 ng/ml/min; P < .05). After 3 hours, the median fetal concentration of rivaroxaban was 69.5 ng/mL (IQR, 67.9–84.7 ng/mL), and the median fetal-to-maternal ratio was 0.69 (IQR, 0.58–0.73). The fetal-to-maternal ratios of rivaroxaban were compared with those of antipyrine ( Figure 2 ). Lines have been fitted through the first few data points (0-60 minutes), because transfer is expected to occur only in the maternal-to-fetal direction. The fetal-to-maternal ratio for antipyrine increased at a slightly faster rate compared with that of rivaroxaban: 0.064 (IQR, 0.0059–0.0067) vs 0.0056 (IQR, 0.0054–0.0058) min –1 ; P = .028].
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
